A research team from Cornell Tech is addressing the lack of technical support for various end-user hardware options, using a tool they developed that renders 3D models via a mobile device.
The tool, called SplatOverflow is intended to use the renders to make hardware issues easier to identify and diagnose. It is based on the similarly named tool StackOverflow, which performs an identical function for software issues.
“As tools for designing and manufacturing hardware become more accessible, smaller producers can develop and distribute novel hardware. However, processes for supporting end-user hardware troubleshooting or routine maintenance aren’t well defined. As a result, providing technical support for hardware remains ad-hoc and challenging to scale,” the team said in its research blurb.
SplatOverflow is the brainchild of Amritansh Kwatra, a Ph.D., an information science fellow at Cornell Tech. The research was directed by Thijs Roumen, assistant professor at Cornell Tech. The research was carried out at Matter of Tech Lab.
What SplatOverflow does is put the design information, documentation, and end-user discussions, in one place, making it easier for people who might not have expertise in hardware, Kwatra told the Cornell Chronicle.
The tool first creates a 3D model of the device that needs troubleshooting, with information about the model included. This allows the specialist and the user to view the scene together to identify the issue. The rendered model is called a “SplatOverflow scene.” It includes two parts– a 3D Gaussian Splat the 3D rendering of the device that needs troubleshooting; and the computer-aided design (CAD) model, of the additional information.
SplatOverflow allows experts to compare the 3D Gaussian Splat with the CAD model, to get a clearer idea of the original design and function. Users reporting a remote issue can point on a virtual 3D model where they believe the issue originates.
In their research, the team demonstrated SplatOverflow troubleshooted issues on various hardware, including a pick-and-place machine, a 3D printer, and an open-source e-reader.
Roumen noted the team is looking into SplatOverflow being used in several use cases, such as machine sanitation, maintenance in agriculture, and education, among others– also noting the tool is available for others to develop within other contexts.
“We are thrilled to explore longer-term deployment, to see how we can gradually grow this into a real, open-source platform for hardware support,” he said.
SplatOverflow and its research will be presented at the Association of Computing Machinery’s CHI conference on April 30, in Yokohama, Japan.